The Potential of Near-infrared Spectroscopy to Predict Soil Nutrient Contents Based on Soil Color

Near-infrared spectroscopy (NIRs) analysis in laboratory-based settings has the potential to predict soil elements. The aim was to explore the effects of soil color on the prediction of total nitrogen (N), available phosphorus (P), and extractable potassium (K) contents using near-infrared spectroscopy in the range of 1000–2500 nm. Two hundred forty soil samples were collected from a paddy field in northeast Thailand. We divided the soil samples based on soil color using the Munsell color chart to construct a model to predict nutrient contents based on soil color. Regression models for soil nutrient contents were developed using partial least squares regression (PLSR) models. The best predictions were obtained for N (R² = 0.87, RMSE = 0.131), P (R² = 0.87, RMSE = 7.713) and K (R² = 0.77, RMSE = 14.944). This research demonstrates the viability of employing Near-Infrared spectroscopy (NIRs) as a reasonable method for predicting soil nutrient contents.